Hybrid approach to approximate string matching using machine learning

Abstract

Systems, apparatuses, and methods are provided for identifying a corresponding string stored in memory based on an incomplete input string. A system can analyze and produce phonetic and distance metrics for a plurality of strings stored in memory by comparing the plurality of strings to an incomplete input string. These similarity metrics can be used as the input to a machine learning model, which can quickly and accurately provide a classification. This classification can be used to identify a string stored in memory that corresponds to the incomplete input string.

Description

BACKGROUND[0001]In recent years searching through databases has become an extremely ubiquitous part of business, government, and industrial operations. Outside of commercial search, most organizations maintain some sort of database which can be searched and accessed for information relevant to business operations, such as the names and phone numbers of employees and clients, sales records, and active projects.[0002]These databases receive inputs in the form of queries, typically taking the form of a “string,” or an ordered set of “characters.” Commonly strings take the form of a name of a person or business. Databases rely on accurate entry of string queries. A database which is queried with an incorrect or incomplete string cannot produce the desired information.[0003]Human error is a common cause of producing incorrect or incomplete strings. Over the course of thousands and thousands of keystrokes, a human typist will produce several incorrect or incomplete input strings. Likewise...

AI Technical Summary

Benefits of technology

[0008]Embodiments of the invention provide for advantages over conventional methods because a string stored in memory can quickly and accurately be determined from an incomplete input string. These and other embodiments of the invention are described in detail below. For example, other embodiments are directed to systems, devices, and computer readable media associated with methods described herein.

Problems solved by technology

A database which is queried with an incorrect or incomplete string cannot produce the desired information.

Human error is a common cause of producing incorrect or incomplete strings.

Likewise, digital transmission error, such as noise, interference or distortion can result in incorrect or incomplete strings.

These errors often result in cost or loss to businesses which rely on searching databases using strings.

However, language and string construction in general is typically very complicated.

As a result rule-based methods are costly to produce and implement.

Images